krak/zstd
1
0
Fork 0
mirror of https://github.com/facebook/zstd.git synced 2025-10-31 16:47:48 +02:00
Code Issues Releases 1 Activity

Merge branch 'dev' into http-to-https

Browse Source
This commit is contained in:
daniellerozenblit
2022-12-15 10:46:13 -05:00
committed by GitHub
parent 4dffc35f2e 728e73ebb4
commit e2fc93340f
34 changed files with 574 additions and 534 deletions
Download Patch File Download Diff File Expand all files Collapse all files

39
lib/legacy/zstd_v01.c
View File

@@ -190,25 +190,6 @@ typedef signed long long S64;
/****************************************************************
* Memory I/O
*****************************************************************/
/* FSE_FORCE_MEMORY_ACCESS
* By default, access to unaligned memory is controlled by `memcpy()`, which is safe and portable.
* Unfortunately, on some target/compiler combinations, the generated assembly is sub-optimal.
* The below switch allow to select different access method for improved performance.
* Method 0 (default) : use `memcpy()`. Safe and portable.
* Method 1 : `__packed` statement. It depends on compiler extension (ie, not portable).
* This method is safe if your compiler supports it, and *generally* as fast or faster than `memcpy`.
* Method 2 : direct access. This method is portable but violate C standard.
* It can generate buggy code on targets generating assembly depending on alignment.
* But in some circumstances, it's the only known way to get the most performance (ie GCC + ARMv6)
* See https://fastcompression.blogspot.fr/2015/08/accessing-unaligned-memory.html for details.
* Prefer these methods in priority order (0 > 1 > 2)
*/
#ifndef FSE_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */
# if defined(__INTEL_COMPILER) || defined(__GNUC__) || defined(__ICCARM__)
# define FSE_FORCE_MEMORY_ACCESS 1
# endif
#endif
static unsigned FSE_32bits(void)
{
@@ -221,24 +202,6 @@ static unsigned FSE_isLittleEndian(void)
return one.c[0];
}
#if defined(FSE_FORCE_MEMORY_ACCESS) && (FSE_FORCE_MEMORY_ACCESS==2)
static U16 FSE_read16(const void* memPtr) { return *(const U16*) memPtr; }
static U32 FSE_read32(const void* memPtr) { return *(const U32*) memPtr; }
static U64 FSE_read64(const void* memPtr) { return *(const U64*) memPtr; }
#elif defined(FSE_FORCE_MEMORY_ACCESS) && (FSE_FORCE_MEMORY_ACCESS==1)
/* __pack instructions are safer, but compiler specific, hence potentially problematic for some compilers */
/* currently only defined for gcc and icc */
typedef union { U16 u16; U32 u32; U64 u64; } __attribute__((packed)) unalign;
static U16 FSE_read16(const void* ptr) { return ((const unalign*)ptr)->u16; }
static U32 FSE_read32(const void* ptr) { return ((const unalign*)ptr)->u32; }
static U64 FSE_read64(const void* ptr) { return ((const unalign*)ptr)->u64; }
#else
static U16 FSE_read16(const void* memPtr)
{
U16 val; memcpy(&val, memPtr, sizeof(val)); return val;
@@ -254,8 +217,6 @@ static U64 FSE_read64(const void* memPtr)
U64 val; memcpy(&val, memPtr, sizeof(val)); return val;
}
#endif /* FSE_FORCE_MEMORY_ACCESS */
static U16 FSE_readLE16(const void* memPtr)
{
if (FSE_isLittleEndian())

48
lib/legacy/zstd_v02.c
View File

@@ -115,24 +115,6 @@ extern "C" {
/****************************************************************
* Memory I/O
*****************************************************************/
/* MEM_FORCE_MEMORY_ACCESS
* By default, access to unaligned memory is controlled by `memcpy()`, which is safe and portable.
* Unfortunately, on some target/compiler combinations, the generated assembly is sub-optimal.
* The below switch allow to select different access method for improved performance.
* Method 0 (default) : use `memcpy()`. Safe and portable.
* Method 1 : `__packed` statement. It depends on compiler extension (ie, not portable).
* This method is safe if your compiler supports it, and *generally* as fast or faster than `memcpy`.
* Method 2 : direct access. This method is portable but violate C standard.
* It can generate buggy code on targets generating assembly depending on alignment.
* But in some circumstances, it's the only known way to get the most performance (ie GCC + ARMv6)
* See https://fastcompression.blogspot.fr/2015/08/accessing-unaligned-memory.html for details.
* Prefer these methods in priority order (0 > 1 > 2)
*/
#ifndef MEM_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */
# if defined(__INTEL_COMPILER) || defined(__GNUC__) || defined(__ICCARM__)
# define MEM_FORCE_MEMORY_ACCESS 1
# endif
#endif
MEM_STATIC unsigned MEM_32bits(void) { return sizeof(void*)==4; }
MEM_STATIC unsigned MEM_64bits(void) { return sizeof(void*)==8; }
@@ -143,33 +125,6 @@ MEM_STATIC unsigned MEM_isLittleEndian(void)
return one.c[0];
}
#if defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==2)
/* violates C standard on structure alignment.
Only use if no other choice to achieve best performance on target platform */
MEM_STATIC U16 MEM_read16(const void* memPtr) { return *(const U16*) memPtr; }
MEM_STATIC U32 MEM_read32(const void* memPtr) { return *(const U32*) memPtr; }
MEM_STATIC U64 MEM_read64(const void* memPtr) { return *(const U64*) memPtr; }
MEM_STATIC void MEM_write16(void* memPtr, U16 value) { *(U16*)memPtr = value; }
#elif defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==1)
/* __pack instructions are safer, but compiler specific, hence potentially problematic for some compilers */
/* currently only defined for gcc and icc */
typedef union { U16 u16; U32 u32; U64 u64; } __attribute__((packed)) unalign;
MEM_STATIC U16 MEM_read16(const void* ptr) { return ((const unalign*)ptr)->u16; }
MEM_STATIC U32 MEM_read32(const void* ptr) { return ((const unalign*)ptr)->u32; }
MEM_STATIC U64 MEM_read64(const void* ptr) { return ((const unalign*)ptr)->u64; }
MEM_STATIC void MEM_write16(void* memPtr, U16 value) { ((unalign*)memPtr)->u16 = value; }
#else
/* default method, safe and standard.
can sometimes prove slower */
MEM_STATIC U16 MEM_read16(const void* memPtr)
{
U16 val; memcpy(&val, memPtr, sizeof(val)); return val;
@@ -190,9 +145,6 @@ MEM_STATIC void MEM_write16(void* memPtr, U16 value)
memcpy(memPtr, &value, sizeof(value));
}
#endif /* MEM_FORCE_MEMORY_ACCESS */
MEM_STATIC U16 MEM_readLE16(const void* memPtr)
{
if (MEM_isLittleEndian())

49
lib/legacy/zstd_v03.c
View File

@@ -116,24 +116,6 @@ extern "C" {
/****************************************************************
* Memory I/O
*****************************************************************/
/* MEM_FORCE_MEMORY_ACCESS
* By default, access to unaligned memory is controlled by `memcpy()`, which is safe and portable.
* Unfortunately, on some target/compiler combinations, the generated assembly is sub-optimal.
* The below switch allow to select different access method for improved performance.
* Method 0 (default) : use `memcpy()`. Safe and portable.
* Method 1 : `__packed` statement. It depends on compiler extension (ie, not portable).
* This method is safe if your compiler supports it, and *generally* as fast or faster than `memcpy`.
* Method 2 : direct access. This method is portable but violate C standard.
* It can generate buggy code on targets generating assembly depending on alignment.
* But in some circumstances, it's the only known way to get the most performance (ie GCC + ARMv6)
* See https://fastcompression.blogspot.fr/2015/08/accessing-unaligned-memory.html for details.
* Prefer these methods in priority order (0 > 1 > 2)
*/
#ifndef MEM_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */
# if defined(__INTEL_COMPILER) || defined(__GNUC__) || defined(__ICCARM__)
# define MEM_FORCE_MEMORY_ACCESS 1
# endif
#endif
MEM_STATIC unsigned MEM_32bits(void) { return sizeof(void*)==4; }
MEM_STATIC unsigned MEM_64bits(void) { return sizeof(void*)==8; }
@@ -144,33 +126,6 @@ MEM_STATIC unsigned MEM_isLittleEndian(void)
return one.c[0];
}
#if defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==2)
/* violates C standard on structure alignment.
Only use if no other choice to achieve best performance on target platform */
MEM_STATIC U16 MEM_read16(const void* memPtr) { return *(const U16*) memPtr; }
MEM_STATIC U32 MEM_read32(const void* memPtr) { return *(const U32*) memPtr; }
MEM_STATIC U64 MEM_read64(const void* memPtr) { return *(const U64*) memPtr; }
MEM_STATIC void MEM_write16(void* memPtr, U16 value) { *(U16*)memPtr = value; }
#elif defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==1)
/* __pack instructions are safer, but compiler specific, hence potentially problematic for some compilers */
/* currently only defined for gcc and icc */
typedef union { U16 u16; U32 u32; U64 u64; } __attribute__((packed)) unalign;
MEM_STATIC U16 MEM_read16(const void* ptr) { return ((const unalign*)ptr)->u16; }
MEM_STATIC U32 MEM_read32(const void* ptr) { return ((const unalign*)ptr)->u32; }
MEM_STATIC U64 MEM_read64(const void* ptr) { return ((const unalign*)ptr)->u64; }
MEM_STATIC void MEM_write16(void* memPtr, U16 value) { ((unalign*)memPtr)->u16 = value; }
#else
/* default method, safe and standard.
can sometimes prove slower */
MEM_STATIC U16 MEM_read16(const void* memPtr)
{
U16 val; memcpy(&val, memPtr, sizeof(val)); return val;
@@ -191,10 +146,6 @@ MEM_STATIC void MEM_write16(void* memPtr, U16 value)
memcpy(memPtr, &value, sizeof(value));
}
#endif /* MEM_FORCE_MEMORY_ACCESS */
MEM_STATIC U16 MEM_readLE16(const void* memPtr)
{
if (MEM_isLittleEndian())

48
lib/legacy/zstd_v04.c
View File

@@ -87,24 +87,6 @@ extern "C" {
/****************************************************************
* Memory I/O
*****************************************************************/
/* MEM_FORCE_MEMORY_ACCESS
* By default, access to unaligned memory is controlled by `memcpy()`, which is safe and portable.
* Unfortunately, on some target/compiler combinations, the generated assembly is sub-optimal.
* The below switch allow to select different access method for improved performance.
* Method 0 (default) : use `memcpy()`. Safe and portable.
* Method 1 : `__packed` statement. It depends on compiler extension (ie, not portable).
* This method is safe if your compiler supports it, and *generally* as fast or faster than `memcpy`.
* Method 2 : direct access. This method is portable but violate C standard.
* It can generate buggy code on targets generating assembly depending on alignment.
* But in some circumstances, it's the only known way to get the most performance (ie GCC + ARMv6)
* See https://fastcompression.blogspot.fr/2015/08/accessing-unaligned-memory.html for details.
* Prefer these methods in priority order (0 > 1 > 2)
*/
#ifndef MEM_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */
# if defined(__INTEL_COMPILER) || defined(__GNUC__) || defined(__ICCARM__)
# define MEM_FORCE_MEMORY_ACCESS 1
# endif
#endif
MEM_STATIC unsigned MEM_32bits(void) { return sizeof(void*)==4; }
MEM_STATIC unsigned MEM_64bits(void) { return sizeof(void*)==8; }
@@ -115,33 +97,6 @@ MEM_STATIC unsigned MEM_isLittleEndian(void)
return one.c[0];
}
#if defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==2)
/* violates C standard on structure alignment.
Only use if no other choice to achieve best performance on target platform */
MEM_STATIC U16 MEM_read16(const void* memPtr) { return *(const U16*) memPtr; }
MEM_STATIC U32 MEM_read32(const void* memPtr) { return *(const U32*) memPtr; }
MEM_STATIC U64 MEM_read64(const void* memPtr) { return *(const U64*) memPtr; }
MEM_STATIC void MEM_write16(void* memPtr, U16 value) { *(U16*)memPtr = value; }
#elif defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==1)
/* __pack instructions are safer, but compiler specific, hence potentially problematic for some compilers */
/* currently only defined for gcc and icc */
typedef union { U16 u16; U32 u32; U64 u64; } __attribute__((packed)) unalign;
MEM_STATIC U16 MEM_read16(const void* ptr) { return ((const unalign*)ptr)->u16; }
MEM_STATIC U32 MEM_read32(const void* ptr) { return ((const unalign*)ptr)->u32; }
MEM_STATIC U64 MEM_read64(const void* ptr) { return ((const unalign*)ptr)->u64; }
MEM_STATIC void MEM_write16(void* memPtr, U16 value) { ((unalign*)memPtr)->u16 = value; }
#else
/* default method, safe and standard.
can sometimes prove slower */
MEM_STATIC U16 MEM_read16(const void* memPtr)
{
U16 val; memcpy(&val, memPtr, sizeof(val)); return val;
@@ -162,9 +117,6 @@ MEM_STATIC void MEM_write16(void* memPtr, U16 value)
memcpy(memPtr, &value, sizeof(value));
}
#endif /* MEM_FORCE_MEMORY_ACCESS */
MEM_STATIC U16 MEM_readLE16(const void* memPtr)
{
if (MEM_isLittleEndian())

52
lib/legacy/zstd_v05.c
View File

@@ -106,24 +106,6 @@ extern "C" {
/*-**************************************************************
* Memory I/O
*****************************************************************/
/* MEM_FORCE_MEMORY_ACCESS :
* By default, access to unaligned memory is controlled by `memcpy()`, which is safe and portable.
* Unfortunately, on some target/compiler combinations, the generated assembly is sub-optimal.
* The below switch allow to select different access method for improved performance.
* Method 0 (default) : use `memcpy()`. Safe and portable.
* Method 1 : `__packed` statement. It depends on compiler extension (ie, not portable).
* This method is safe if your compiler supports it, and *generally* as fast or faster than `memcpy`.
* Method 2 : direct access. This method is portable but violate C standard.
* It can generate buggy code on targets depending on alignment.
* In some circumstances, it's the only known way to get the most performance (ie GCC + ARMv6)
* See https://fastcompression.blogspot.fr/2015/08/accessing-unaligned-memory.html for details.
* Prefer these methods in priority order (0 > 1 > 2)
*/
#ifndef MEM_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */
# if defined(__INTEL_COMPILER) || defined(__GNUC__) || defined(__ICCARM__)
# define MEM_FORCE_MEMORY_ACCESS 1
# endif
#endif
MEM_STATIC unsigned MEM_32bits(void) { return sizeof(void*)==4; }
MEM_STATIC unsigned MEM_64bits(void) { return sizeof(void*)==8; }
@@ -134,37 +116,6 @@ MEM_STATIC unsigned MEM_isLittleEndian(void)
return one.c[0];
}
#if defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==2)
/* violates C standard, by lying on structure alignment.
Only use if no other choice to achieve best performance on target platform */
MEM_STATIC U16 MEM_read16(const void* memPtr) { return *(const U16*) memPtr; }
MEM_STATIC U32 MEM_read32(const void* memPtr) { return *(const U32*) memPtr; }
MEM_STATIC U64 MEM_read64(const void* memPtr) { return *(const U64*) memPtr; }
MEM_STATIC void MEM_write16(void* memPtr, U16 value) { *(U16*)memPtr = value; }
MEM_STATIC void MEM_write32(void* memPtr, U32 value) { *(U32*)memPtr = value; }
MEM_STATIC void MEM_write64(void* memPtr, U64 value) { *(U64*)memPtr = value; }
#elif defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==1)
/* __pack instructions are safer, but compiler specific, hence potentially problematic for some compilers */
/* currently only defined for gcc and icc */
typedef union { U16 u16; U32 u32; U64 u64; size_t st; } __attribute__((packed)) unalign;
MEM_STATIC U16 MEM_read16(const void* ptr) { return ((const unalign*)ptr)->u16; }
MEM_STATIC U32 MEM_read32(const void* ptr) { return ((const unalign*)ptr)->u32; }
MEM_STATIC U64 MEM_read64(const void* ptr) { return ((const unalign*)ptr)->u64; }
MEM_STATIC void MEM_write16(void* memPtr, U16 value) { ((unalign*)memPtr)->u16 = value; }
MEM_STATIC void MEM_write32(void* memPtr, U32 value) { ((unalign*)memPtr)->u32 = value; }
MEM_STATIC void MEM_write64(void* memPtr, U64 value) { ((unalign*)memPtr)->u64 = value; }
#else
/* default method, safe and standard.
can sometimes prove slower */
MEM_STATIC U16 MEM_read16(const void* memPtr)
{
U16 val; memcpy(&val, memPtr, sizeof(val)); return val;
@@ -195,9 +146,6 @@ MEM_STATIC void MEM_write64(void* memPtr, U64 value)
memcpy(memPtr, &value, sizeof(value));
}
#endif /* MEM_FORCE_MEMORY_ACCESS */
MEM_STATIC U16 MEM_readLE16(const void* memPtr)
{
if (MEM_isLittleEndian())

51
lib/legacy/zstd_v06.c
View File

@@ -108,24 +108,6 @@ extern "C" {
/*-**************************************************************
* Memory I/O
*****************************************************************/
/* MEM_FORCE_MEMORY_ACCESS :
* By default, access to unaligned memory is controlled by `memcpy()`, which is safe and portable.
* Unfortunately, on some target/compiler combinations, the generated assembly is sub-optimal.
* The below switch allow to select different access method for improved performance.
* Method 0 (default) : use `memcpy()`. Safe and portable.
* Method 1 : `__packed` statement. It depends on compiler extension (ie, not portable).
* This method is safe if your compiler supports it, and *generally* as fast or faster than `memcpy`.
* Method 2 : direct access. This method is portable but violate C standard.
* It can generate buggy code on targets depending on alignment.
* In some circumstances, it's the only known way to get the most performance (ie GCC + ARMv6)
* See https://fastcompression.blogspot.fr/2015/08/accessing-unaligned-memory.html for details.
* Prefer these methods in priority order (0 > 1 > 2)
*/
#ifndef MEM_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */
# if defined(__INTEL_COMPILER) || defined(__GNUC__) || defined(__ICCARM__)
# define MEM_FORCE_MEMORY_ACCESS 1
# endif
#endif
MEM_STATIC unsigned MEM_32bits(void) { return sizeof(size_t)==4; }
MEM_STATIC unsigned MEM_64bits(void) { return sizeof(size_t)==8; }
@@ -136,33 +118,6 @@ MEM_STATIC unsigned MEM_isLittleEndian(void)
return one.c[0];
}
#if defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==2)
/* violates C standard, by lying on structure alignment.
Only use if no other choice to achieve best performance on target platform */
MEM_STATIC U16 MEM_read16(const void* memPtr) { return *(const U16*) memPtr; }
MEM_STATIC U32 MEM_read32(const void* memPtr) { return *(const U32*) memPtr; }
MEM_STATIC U64 MEM_read64(const void* memPtr) { return *(const U64*) memPtr; }
MEM_STATIC void MEM_write16(void* memPtr, U16 value) { *(U16*)memPtr = value; }
#elif defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==1)
/* __pack instructions are safer, but compiler specific, hence potentially problematic for some compilers */
/* currently only defined for gcc and icc */
typedef union { U16 u16; U32 u32; U64 u64; size_t st; } __attribute__((packed)) unalign;
MEM_STATIC U16 MEM_read16(const void* ptr) { return ((const unalign*)ptr)->u16; }
MEM_STATIC U32 MEM_read32(const void* ptr) { return ((const unalign*)ptr)->u32; }
MEM_STATIC U64 MEM_read64(const void* ptr) { return ((const unalign*)ptr)->u64; }
MEM_STATIC void MEM_write16(void* memPtr, U16 value) { ((unalign*)memPtr)->u16 = value; }
#else
/* default method, safe and standard.
can sometimes prove slower */
MEM_STATIC U16 MEM_read16(const void* memPtr)
{
U16 val; memcpy(&val, memPtr, sizeof(val)); return val;
@@ -183,9 +138,6 @@ MEM_STATIC void MEM_write16(void* memPtr, U16 value)
memcpy(memPtr, &value, sizeof(value));
}
#endif /* MEM_FORCE_MEMORY_ACCESS */
MEM_STATIC U32 MEM_swap32(U32 in)
{
#if defined(_MSC_VER) /* Visual Studio */
@@ -4035,7 +3987,8 @@ size_t ZBUFFv06_decompressContinue(ZBUFFv06_DCtx* zbd,
size_t const toLoad = hSize - zbd->lhSize; /* if hSize!=0, hSize > zbd->lhSize */
if (ZSTDv06_isError(hSize)) return hSize;
if (toLoad > (size_t)(iend-ip)) { /* not enough input to load full header */
memcpy(zbd->headerBuffer + zbd->lhSize, ip, iend-ip);
if (ip != NULL)
memcpy(zbd->headerBuffer + zbd->lhSize, ip, iend-ip);
zbd->lhSize += iend-ip;
*dstCapacityPtr = 0;
return (hSize - zbd->lhSize) + ZSTDv06_blockHeaderSize; /* remaining header bytes + next block header */

50
lib/legacy/zstd_v07.c
View File

@@ -268,24 +268,6 @@ extern "C" {
/*-**************************************************************
* Memory I/O
*****************************************************************/
/* MEM_FORCE_MEMORY_ACCESS :
* By default, access to unaligned memory is controlled by `memcpy()`, which is safe and portable.
* Unfortunately, on some target/compiler combinations, the generated assembly is sub-optimal.
* The below switch allow to select different access method for improved performance.
* Method 0 (default) : use `memcpy()`. Safe and portable.
* Method 1 : `__packed` statement. It depends on compiler extension (ie, not portable).
* This method is safe if your compiler supports it, and *generally* as fast or faster than `memcpy`.
* Method 2 : direct access. This method is portable but violate C standard.
* It can generate buggy code on targets depending on alignment.
* In some circumstances, it's the only known way to get the most performance (ie GCC + ARMv6)
* See https://fastcompression.blogspot.fr/2015/08/accessing-unaligned-memory.html for details.
* Prefer these methods in priority order (0 > 1 > 2)
*/
#ifndef MEM_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */
# if defined(__INTEL_COMPILER) || defined(__GNUC__) || defined(__ICCARM__)
# define MEM_FORCE_MEMORY_ACCESS 1
# endif
#endif
MEM_STATIC unsigned MEM_32bits(void) { return sizeof(size_t)==4; }
MEM_STATIC unsigned MEM_64bits(void) { return sizeof(size_t)==8; }
@@ -296,33 +278,6 @@ MEM_STATIC unsigned MEM_isLittleEndian(void)
return one.c[0];
}
#if defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==2)
/* violates C standard, by lying on structure alignment.
Only use if no other choice to achieve best performance on target platform */
MEM_STATIC U16 MEM_read16(const void* memPtr) { return *(const U16*) memPtr; }
MEM_STATIC U32 MEM_read32(const void* memPtr) { return *(const U32*) memPtr; }
MEM_STATIC U64 MEM_read64(const void* memPtr) { return *(const U64*) memPtr; }
MEM_STATIC void MEM_write16(void* memPtr, U16 value) { *(U16*)memPtr = value; }
#elif defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==1)
/* __pack instructions are safer, but compiler specific, hence potentially problematic for some compilers */
/* currently only defined for gcc and icc */
typedef union { U16 u16; U32 u32; U64 u64; size_t st; } __attribute__((packed)) unalign;
MEM_STATIC U16 MEM_read16(const void* ptr) { return ((const unalign*)ptr)->u16; }
MEM_STATIC U32 MEM_read32(const void* ptr) { return ((const unalign*)ptr)->u32; }
MEM_STATIC U64 MEM_read64(const void* ptr) { return ((const unalign*)ptr)->u64; }
MEM_STATIC void MEM_write16(void* memPtr, U16 value) { ((unalign*)memPtr)->u16 = value; }
#else
/* default method, safe and standard.
can sometimes prove slower */
MEM_STATIC U16 MEM_read16(const void* memPtr)
{
U16 val; memcpy(&val, memPtr, sizeof(val)); return val;
@@ -343,8 +298,6 @@ MEM_STATIC void MEM_write16(void* memPtr, U16 value)
memcpy(memPtr, &value, sizeof(value));
}
#endif /* MEM_FORCE_MEMORY_ACCESS */
MEM_STATIC U32 MEM_swap32(U32 in)
{
#if defined(_MSC_VER) /* Visual Studio */
@@ -4417,7 +4370,8 @@ size_t ZBUFFv07_decompressContinue(ZBUFFv07_DCtx* zbd,
if (hSize != 0) {
size_t const toLoad = hSize - zbd->lhSize; /* if hSize!=0, hSize > zbd->lhSize */
if (toLoad > (size_t)(iend-ip)) { /* not enough input to load full header */
memcpy(zbd->headerBuffer + zbd->lhSize, ip, iend-ip);
if (ip != NULL)
memcpy(zbd->headerBuffer + zbd->lhSize, ip, iend-ip);
zbd->lhSize += iend-ip;
*dstCapacityPtr = 0;
return (hSize - zbd->lhSize) + ZSTDv07_blockHeaderSize; /* remaining header bytes + next block header */